Young - Computational chemistry

A.4 MOLECULAR MECHANICS 347

Contact information: Chemical Computing Group Inc. 1010 Sherbrooke Street, Suite 910

Montreal, Quebec, Canada H3A 2R7 [phone] http://www.chemcomp.com/ [e-mail]

A.4.3 PC Model

PC Model (we tested Version 7.00) is a molecular mechanics program with a graphic user interface. It can also be used as a graphic interface for AMPAC, MOPAC, and Gaussian. Separately sold modules are available for displaying orbitals and vibrational modes for these programs as well as Hondo and Jaguar. It is able to read and write a variety of chemical-structure data ®le formats. The native ®le format can hold many structures that can be used for either batch processing or displaying an animation.

The available force ®elds are MMX, MM3, and MMFF94. Some of the supported calculations are geometry optimization, molecular dynamics, and a simulated annealing docking algorithm. There is also an automated algorithm for computing rotational energy barriers. A p system semiempirical calculation can be incorporated for modeling aromatic molecules.

PC Model has some features that are not found in many other molecular mechanics programs. This is one of the few programs that outputs the energy given by the force ®eld and the heat of formation and a strain energy. Atom types for describing transition structures in the MMX force ®eld are included. There is a metal coordination option for setting up calculations with metal atoms. There are also molecular similarity and conformation search functions.

The graphic interface can be used to build structures in either a twodimensional drawing mode or a three-dimensional building mode. There are separate builders for amino acids, sugars, and nucleic acids. The amino acid and sugar builders would connect the units to form a chain. The nucleic acid builder placed the nucleotides on the screen, but did not connect them into a strand. There are also libraries of prede®ned ring groups, transition-state geometries, organic functional groups, and organometallic groups. The reviewer felt that the builder was reasonably easy to use.

Molecules can be rendered as stick ®gures, ball and stick, CPK, and ribbons. Dot surfaces can also be included. Some regions were incorrectly shaded for small molecules on our test system running at 800 600 resolution with 24-bit color. The display uses a black background, but graphics are saved or printed with a white background. Overall, the rendering is adequate.

The user manual includes both a function reference and some short tutorials. There is also a tutorial in the following reference: M. F. Schlecht (1998). Molecular Modeling on the PC. New York: John Wiley & Sons.

348 APPENDIX A SOFTWARE PACKAGES

Price category: production

Platforms: PC (Windows and Linux), SGI Contact information: Serena Software

Box 3076

Bloomington, IN 47402-3076 [phone] http://www.serenasoft.com/ [e-mail]

A.4.4 TINKER

TINKER (we tested Version 3.7) is a collection of programs for performing molecular mechanics and dynamics calculations. All the executables read the same ASCII input ®les. TINKER does not come with a graphic interface, although it can be used with RasMol, ChemDraw, Chem3D, gOpenMol, MOLDEN, and ReView. Protein database ®les can be converted to TINKER ®les, making it possible to import molecular structures from many sources. Output can be generated in a format displayed with Sybyl, Insight II, or Xmol. Source code is available.

The force ®elds currently available are AMBER-95, CHARMM22, MM2(91), MM3(99), OPLS-AA, OPLS-UA, and TINKER. TINKER can be used for geometry optimization, molecular dynamics, simulated annealing, normal vibrational mode analysis, conformation searching, distance-geometry, free energy perturbation, and ®nding conformational transition states. Fractional coordinates may then be used for importing molecular structures and hence the parameterization of force ®eld terms. Molecular structures can also be generated from a peptide sequence. In addition, molecular properties, such as energy analysis and molecular volumes and surface areas, can be computed.

The molecular structure input requires atom types to be assigned, which are not the same from one force ®eld to the next. The input also includes a list of bonds in the molecule. There is not a module to automatically assign atom types. Most of the modules use a Cartesian coordinate molecular structure, except for a few that work with torsional space. The same keyword ®le is read by all the executables. A little bit of input is obtained by the program either interactively or from an ASCII ®le piped to standard input, which makes for a somewhat cryptic input ®le. This system of common input ®les and the user choosing which executables to run give TINKER the ability to run very sophisticated simulations while keeping the input required for simple calculations fairly minimal within the limitations mentioned here.

The TINKER documentation provides a description of the input, but not a tutorial. Documentation is available as html, Acrobat, or postscript. A set of example input ®les is provided. The researcher can expect to invest some time in learning to use this system of programs. Most of the executables seem to be fairly robust and as tolerant as possible of variations in the input format. When

A.5 GRAPHICS PACKAGES

349

possible, default values are assumed to minimize the amount of input information that must be provided. TINKER is well suited for researchers wishing to call it from their own front-end programs.

Price category: free

Platforms: PC (Windows and Linux), SGI, Macintosh, RS/6000, Alpha, HP-UX, Sun

Contact information: Jay W. Ponder Biochemistry, Box 8231

Washington University Medical School 660 South Euclid Ave.

St. Lewis, MO 63110 [phone]

http://dasher.wustl.edu/tinker/

[e-mail]

A.5 GRAPHICS PACKAGES

The programs discussed in this section are designed not for running calculations, but as graphic interfaces for constructing input ®les and/or viewing results. Some more general-purpose programs were omitted in favor of software designed speci®cally for chemistry. Many other chemistry visualizaton programs were omitted because they had narrow applicability or are less widely used. These capabilities are also integrated with computational programs in many of the packages discussed previously in this appendix.

A.5.1 GaussView

GaussView (we used Version 2.08) is a graphic interface for use with the Gaussian ab initio program. It can be used to build molecules, set up the options in the input ®le, run a calculation, and display results. GaussView uses the molecule builder that was written by SemiChem, but has screens for setting up calculations that are di¨erent from those in the AMPAC GUI sold by SemiChem.

The program has several building modes. Compounds can be built one atom at a time by selecting the element and hybridization. There are also libraries of ring systems, amino acids, nucleosides, and common organic functional groups. The user can manually set bond lengths, angles, and dihedral angles. There is a clean function that gives an initial optimization of the structure using a rulebased VSEPR algorithm. There is a Z-matrix editor that gives some control over how the Z-matrix is constructed, but does not go as far as giving the user the ability to enforce symmetry constraints. GaussView can also be used to set

350 APPENDIX A SOFTWARE PACKAGES

up ONIOM QM/MM calculations. Our reviewer felt that the graphic moleculebuilding functions were very easy to use.

There is a screen to set up the calculation that has menus for the most widely used functions. Many users will still need to know many of the keywords, which can be typed in. There was no default comment statement, so the input ®le created would not be valid if the user forgot to include a comment. A calculation can be started from the graphic interface, which will be run interactively by default. The script that launches the calculation was not too di½cult to modify for use with a job-queueing system.

The molecular structures were rendered with good-quality shading on a blue background. Isosurfaces produced from cube ®les or checkpoint ®les also looked nice. Molecular vibrations can be animated on screen and vibrational displacement vectors displayed. The vibrational line spectrum may be displayed too, but the user has no control over the axes. There is no way to set the background color. The display can be saved using several image ®le formats.

Price category: production, departmental Platforms: RS/6000, SGI, Alpha, PC (Windows) Contact information: Gaussian, Inc.

Carnegie O½ce Park, Bldg. 6 Suite 230

Pittsburgh, PA 15106 [phone] http://www.gaussian.com/ [e-mail]

A.5.2 Molden

Molden (we tested Version 3.6) is a molecular display program. It can display molecular geometries read from a number of molecular ®le formats. Various views of the wave function can be displayed from the output of the Gaussian and GAMESS programs. Some functionality is available from MOPAC and AMPAC ®les. Conversion programs are available to import wave functions from ADF, MOLPRO, ACES II, MOLCAS, DALTON, Jaguar, and HONDO.

Molden can display molecular geometries in a variety of formats, including lines, tubes, ball and stick, ribbons, and CPK. The user has some control over colors and sizes. Molden also has features designed for the display of proteins and crystal structures. The display can be exported as postscript, VRML, Povray, and image ®les. It can also be con®gured as a chemical mime viewer.

The program has a Z-matrix editor, which is not the same as a graphic molecule builder. This allows the user to display the Z-matrix and then de®ne

A.5 GRAPHICS PACKAGES

351

which parameters are ®xed or equivalent to other parameters. Molden can be con®gured to optimize the geometry by passing data to the TINKER program. Molecular structures can be exported in formats compatible with many popular software programs.

Wave functions can be visualized as the total electron density, orbital densities, electrostatic potential, atomic densities, or the Laplacian of the electron density. The program computes the data from the basis functions and molecular orbital coe½cients. Thus, it does not need a large amount of disk space to store data, but the computation can be time-consuming. Molden can also compute electrostatic charges from the wave function. Several visualization modes are available, including contour plots, three-dimensional isosurfaces, and data slices.

Molden now has the capacity to export a series of GIF ®les, creating an additional ®le with each screen update. Third-party utilities can be used to combine these images in animations with various ®le formats, including ¯i and animated GIF ®les. This is one of the easiest ways to create animations of chemical systems.

Program documentation is available as postscript, text, or Web pages. None of these seemed to give a comprehensive description of the program functionality and controls.

Price category: free, production

Platforms: PC (Windows, Linux, free BSD), SGI, RS/6000, Alpha, Cray, HP-UX, Sun, open VMS, OS2

Contact information: Gijs Schaftenaar CMBI

University of Nijmegen Toernooiveld 1

6525 ED NIJMEGEN, The Netherlands

‡31 24 365 33 69 http://www.cmbi.kun.nl/@schaft/molden/molden.html [e-mail]

A.5.3 WebLab Viewer

WebLab Viewer is a molecular graphics program. We tested two versions: WebLab9 ViewerLiteTM (we tested Version 3.2), a free molecular display program, and WebLab9 ViewerProTM (we tested Version 3.5) for molecule building and display. ViewerPro can also be used as a graphic interface to MedChem Explorer (for drug re®nement), Diversity Explorer (in combinatorial chemistry), and Gene Explorer (in bioinformatics).

Both versions of the WebLab Viewer use a native ®le format that is capable

352 APPENDIX A SOFTWARE PACKAGES

of describing molecular or crystal structures along with surfaces and labels. They also read a dozen di¨erent common molecular ®le formats. Two-dimensional structures can be automatically converted to three-dimensions when the ®le is read. Work can be saved in common molecule ®le formats, VRML, SMILES, and bit-mapped graphic ®les.

WebLab Viewer gives a very-high-quality display suitable for publication and presentation. Molecules can be displayed as lines, sticks, ball and stick, CPK, and polyhedrons. In addition, di¨erent atoms within the same structure may be displayed in di¨erent ways. Text can be added to the display as well as labeling parts of the structure in a variety of ways. The user has control over colors, radii, and display quality. The program can also replicate a unit cell to display a crystal structure. Several types of molecular surfaces can be displayed.

WebLab ViewerPro has all the functionality of ViewerLite plus a number of additional features. ViewerPro can be used to build molecular structures. It has building modes for adding individual atoms, creating chains, and creating rings. All these create structures of carbon atoms. Once the backbone has been created, atoms can be changed to other elements and hydrogens added. This builder worked reasonably well for organic molecules, but seemed somewhat inconvenient for building inorganic structures. There is a function to clean up the shape of the molecule, which does a basic molecular mechanics minimization using a simple Dreiding-type force ®eld. ViewerPro can be used to create animations and has a scripting language to automate tasks.

Price category: free, individual, production Platforms: PC, Mac

Contact information: Molecular Simulations, Inc. 9685 Scranton Road

San Diego, CA 92121-3752 [phone] http://www.msi.com/viewer

A.6 SPECIAL-PURPOSE PROGRAMS

A.6.1 Babel

Babel (we tested Version 1.6) is a utility for converting computational chemistry input ®les from one format to another. It is able to interconvert about 50 different ®le formats, including conversions between SMILES, Cartesian coordinate, and Z-matrix input. The algorithm that generates a Z-matrix from Cartesian coordinates is fairly simplistic, so the Z-matrix will correctly represent the geometry, but will not include symmetry, dummy atoms, and the like. Babel can be run with command line options or in a menu-driven mode. There have been some third-party graphic interfaces created for Babel.

A.6 SPECIAL-PURPOSE PROGRAMS 353

Price category: free

Platforms: PC (Windows and Linux), Unix, Macintosh

Contact information: http://www.ccl.net/pub/chemistry/software/UNIX/babel/

A.6.2 CHEOPS

CHEOPS (we tested Version 3.0.1) is a program for predicting polymer properties. It consists of two programs: The analysis program allows the user to draw the repeat unit structure and will then compute a whole list of properties; the synthesis program allows the user to specify a class of polymers and desired properties and will then try the various permutations of the functional groups to ®nd ones that ®t the requirements. On a Pentium Pro 200 system, the analysis computations were essentially instantaneous and the synthesis computations could take up to a few minutes. There was no automated way to transfer information between the two programs.

CHEOPS is based on the method of atomic constants, which uses atom contributions and an anharmonic oscillator model. Unlike other similar programs, this allows the prediction of polymer network and copolymer properties. A list of 39 properties could be computed. These include permeability, solubility, thermodynamic, microscopic, physical and optical properties. It also predicts the temperature dependence of some of the properties. The program supports common organic functionality as well as halides, As, B, P, Pb, S, Si, and Sn. Files can be saved with individual structures or a database of structures.

The program is very easy to use. The help screens give step-by-step directions for various operations, which are complete but somewhat di½cult to read because of poor English grammar. Additional information on the website is more readable. The synthesis program works well, although it is limited to seven classes of polymers.

Price category: departmental, institutional (initial purchase and annual license fee)

Platforms: PC

Contact information: MillionZillion Software 3306 Decatur Lane

Minneapolis, MN 55426 [phone]

http://www.millionzillion.com/cheops

[e-mail]

A.6.3 CODESSA

CODESSA (we tested Version 2.6) stands for comprehensive descriptors for structural and statistical analysis. It is a conventional QSAR/QSPR program.

354 APPENDIX A SOFTWARE PACKAGES

CODESSA reads molecular structure ®les or output ®les created by other software packages as the starting point for QSAR analysis. It can import computational results from AMPAC, MOPAC, and Gaussian as well as structures in a number of common formats.

CODESSA can compute or import over 500 molecular descriptors. These can be categorized into constitutional, topological, geometric, electrostatic, quantum chemical, and thermodynamic descriptors. There are automated procedures that will omit missing or bad descriptors. Alternatively, the user can manually de®ne any subset of structures or descriptors to be used.

The program incorporates several very automated procedures for choosing and testing possible QSAR equations. These procedures incorporate correlation and intercorrelation coe½cients to ®nd an equation with the best ®t using a minimal number of descriptors. These automated procedures performed very well when creating an equation to predict normal boiling points using a test set that was constructed by our reviewer. There are both statistical and graphic tools, which also makes this package an excellent choice for experts desiring manual control over the process. The QSAR equations obtained are multilinear.

Price category: production, departmental, institutional

Platforms: PC (Windows and Linux), SGI, RS6000, Alpha, Sun, HP-UX Contact information: Semichem

P.O. Box 1649

Shawnee Mission, KS 66222 [phone] http://www.semichem.com/ [e-mail]

A.6.4 gNMR

gNMR (we tested Version 4.0.1) is a program for NMR spectral prediction and simulation. The simulation portion of the program draws the spectrum once the user has input the chemical shifts and coupling constants. gNMR can simulate spectra for any active nuclei, but can predict chemical shifts only for 1H and 13C. The computed spectrum can be compared to experimental data. Our review will only cover the prediction features pertinent to the discussion in Chapter 31.

gNMR can predict 1H and 13C chemical shifts and coupling constants for up to 23 active nuclei (increasing to 49 nuclei in Version 4.1). It uses additivity rules to predict chemical shifts. The computation time is negligible even on lowend microcomputers. The computed shifts are put in a tabular format. The user can click on atoms in the structure display in order to jump to the corresponding row of numerical data.

The program was made somewhat less convenient to use by the fact that it does not have a molecule builder. In order to predict chemical shifts, the molecular structure must be built with some other software package and then im-

A.6 SPECIAL-PURPOSE PROGRAMS 355

ported into gNMR. Structures can be imported in formats produced by a number of popular chemical drawing and modeling programs.

Price category: individual, production Platforms: PC

Contact information: Cherwell Scienti®c Publishing The Magdalen Centre

Oxford Science Park Oxford, OX4 4GA UK

‡44 (0)1865 784800 http://gNMR.cherwell.com/ [e-mail]

A.6.5 MedChem Explorer

WebLab9 MedChem ExplorerTM (we tested Version 1.6) is a drug re®nement package designed for researchers who do not specialize in computational chemistry. It works as a client±server system so that all functionality is available from a PC. WebLab9 ViewerProTM integrates with MedChem Explorer for molecule building and display. A Web-enabled system is then used to submit the calculations to a server.

The functionality available in MedChem Explorer is broken down into a list of available computational experiments, including activity prediction, align/ pharmacophore, overlay molecules, conformer generation, property calculation, and database access. Within each experiment, the Web system walks the user through a series of questions that must be answered sequentially. The task is then submitted to a remote server, where it is performed. The user can view the progress of the work in their Web browser at any time. Once complete, the results of the calculation are stored on the server. The user can then run subsequent experiments starting with those results. The Web interface includes links to help pages at every step of the process.

Activity prediction is based on a list of models (i.e., QSAR models, pharmacophore models, etc.) that are maintained on the server. There is a second level of access so that only authorized users may be allowed to add or delete model entries.

The align/pharmacophore experiment orients the molecules to obtain maximum similarity in chemical features. This application can then generate a pharmacophore model consistent with all the molecules.

The molecular overlay experiment orients the molecules to ®nd the best RMS or ®eld ®t. The ®eld ®t is based on electrostatic and steric interactions. The application can ®nd either the best total alignment of all molecules or the best match of all molecules to a speci®ed target molecule. Alignment can include a database search for conformers that show the best alignment based on the molecules under study.

356 APPENDIX A SOFTWARE PACKAGES

Conformer generation is used to obtain a list of likely conformers of the molecule. This list can include a set number of the lowest-energy conformers or a number of conformers that give the most diversity of possible shapes.

The property calculation experiment o¨ers a list of 34 molecular properties, including thermodynamic, electrostatic, graph theory, geometric properties, and Lipinski properties. These properties are useful for traditional QSAR activity prediction. Some are computed with MOPAC; others are displayed in the browser without units. A table of computed properties can be exported to a Microsoft Excel spreadsheet.

Database access is used to search external databases for molecules most similar to a speci®ed target molecule. MedChem Explorer o¨ers the following databases: ACD, BioByte, National Cancer Institute, Derwent Drug Index, Maybridge, MDL ISISTM, and Daylight, as well as any in-house or corporate databases that the user may have in any of the ISIS, Catalyst, or Daylight formats. The software is con®gured to link to the in-house informatics environment during installation. The user can search with queries based on shape, topology, substructure, or property information.

Overall, WebLab MedChem Explorer is very easy to use. The stepwise job setup works well, assuming that all users will be following a conventional drug re®nement process. It is not a program that can be used for complex simulations requiring the researcher to manually control many details of the simulation.

Price category: contact Client platforms: PC

Contact information: Molecular Simulations, Inc. 9685 Scranton Road

San Diego, CA 92121-3752 [phone] http://www.msi.com/

A.6.6 POLYRATE

POLYRATE (we tested Version 8.0) is a program for computing chemical reaction rates. The MORATE, GAUSSRATE, and AMSOLRATE programs are derived from POLYRATE and designed to work with the MOPAC, GAUSSIAN, and AMSOL programs, respectively.

POLYRATE can be used for computing reaction rates from either the output of electronic structure calculations or using an analytic potential energy surface. If an analytic potential energy surface is used, the user must create subroutines to evaluate the potential energy and its derivatives then relink the program. POLYRATE can be used for unimolecular gas-phase reactions, bimolecular gas-phase reactions, or the reaction of a gas-phase molecule or adsorbed molecule on a solid surface.

The input to POLYRATE is a free-format ASCII ®le. There are a large